@@ -414,6 +414,9 @@ void __init hyperv_init(void)
register_syscore_ops(&hv_syscore_ops);
hyperv_init_cpuhp = cpuhp;
+
+ /* Query the VMs extended capability once, so that it can be cached. */
+ hv_query_ext_cap(0);
return;
remove_cpuhp_state:
@@ -514,3 +517,50 @@ bool hv_is_hibernation_supported(void)
return acpi_sleep_state_supported(ACPI_STATE_S4);
}
EXPORT_SYMBOL_GPL(hv_is_hibernation_supported);
+
+/* Bit mask of the extended capability to query: see HV_EXT_CAPABILITY_xxx */
+bool hv_query_ext_cap(u64 cap_query)
+{
+ /*
+ * The address of the 'hv_extended_cap' variable will be used as an
+ * output parameter to the hypercall below and so it should be
+ * compatible with 'virt_to_phys'. Which means, it's address should be
+ * directly mapped. Use 'static' to keep it compatible; stack variables
+ * can be virtually mapped, making them imcompatible with
+ * 'virt_to_phys'.
+ * Hypercall input/output addresses should also be 8-byte aligned.
+ */
+ static u64 hv_extended_cap __aligned(8);
+ static bool hv_extended_cap_queried;
+ u64 status;
+
+ /*
+ * Querying extended capabilities is an extended hypercall. Check if the
+ * partition supports extended hypercall, first.
+ */
+ if (!(ms_hyperv.priv_high & HV_ENABLE_EXTENDED_HYPERCALLS))
+ return false;
+
+ /* Extended capabilities do not change at runtime. */
+ if (hv_extended_cap_queried)
+ return hv_extended_cap & cap_query;
+
+ status = hv_do_hypercall(HV_EXT_CALL_QUERY_CAPABILITIES, NULL,
+ &hv_extended_cap);
+
+ /*
+ * The query extended capabilities hypercall should not fail under
+ * any normal circumstances. Avoid repeatedly making the hypercall, on
+ * error.
+ */
+ hv_extended_cap_queried = true;
+ status &= HV_HYPERCALL_RESULT_MASK;
+ if (status != HV_STATUS_SUCCESS) {
+ pr_err("Hyper-V: Extended query capabilities hypercall failed 0x%llx\n",
+ status);
+ return false;
+ }
+
+ return hv_extended_cap & cap_query;
+}
+EXPORT_SYMBOL_GPL(hv_query_ext_cap);
@@ -238,11 +238,13 @@ static void __init ms_hyperv_init_platform(void)
* Extract the features and hints
*/
ms_hyperv.features = cpuid_eax(HYPERV_CPUID_FEATURES);
+ ms_hyperv.priv_high = cpuid_ebx(HYPERV_CPUID_FEATURES);
ms_hyperv.misc_features = cpuid_edx(HYPERV_CPUID_FEATURES);
ms_hyperv.hints = cpuid_eax(HYPERV_CPUID_ENLIGHTMENT_INFO);
- pr_info("Hyper-V: features 0x%x, hints 0x%x, misc 0x%x\n",
- ms_hyperv.features, ms_hyperv.hints, ms_hyperv.misc_features);
+ pr_info("Hyper-V: privilege flags low 0x%x, high 0x%x, hints 0x%x, misc 0x%x\n",
+ ms_hyperv.features, ms_hyperv.priv_high, ms_hyperv.hints,
+ ms_hyperv.misc_features);
ms_hyperv.max_vp_index = cpuid_eax(HYPERV_CPUID_IMPLEMENT_LIMITS);
ms_hyperv.max_lp_index = cpuid_ebx(HYPERV_CPUID_IMPLEMENT_LIMITS);
@@ -23,6 +23,7 @@ config HYPERV_UTILS
config HYPERV_BALLOON
tristate "Microsoft Hyper-V Balloon driver"
depends on HYPERV
+ select PAGE_REPORTING
help
Select this option to enable Hyper-V Balloon driver.
@@ -21,6 +21,7 @@
#include <linux/memory.h>
#include <linux/notifier.h>
#include <linux/percpu_counter.h>
+#include <linux/page_reporting.h>
#include <linux/hyperv.h>
#include <asm/hyperv-tlfs.h>
@@ -565,6 +566,8 @@ struct hv_dynmem_device {
* The negotiated version agreed by host.
*/
__u32 version;
+
+ struct page_reporting_dev_info pr_dev_info;
};
static struct hv_dynmem_device dm_device;
@@ -1570,6 +1573,89 @@ static void balloon_onchannelcallback(void *context)
}
+/* Hyper-V only supports reporting 2MB pages or higher */
+#define HV_MIN_PAGE_REPORTING_ORDER 9
+#define HV_MIN_PAGE_REPORTING_LEN (HV_HYP_PAGE_SIZE << HV_MIN_PAGE_REPORTING_ORDER)
+static int hv_free_page_report(struct page_reporting_dev_info *pr_dev_info,
+ struct scatterlist *sgl, unsigned int nents)
+{
+ unsigned long flags;
+ struct hv_memory_hint *hint;
+ int i;
+ u64 status;
+ struct scatterlist *sg;
+
+ WARN_ON_ONCE(nents > HV_MEMORY_HINT_MAX_GPA_PAGE_RANGES);
+ WARN_ON_ONCE(sgl->length < HV_MIN_PAGE_REPORTING_LEN);
+ local_irq_save(flags);
+ hint = *(struct hv_memory_hint **)this_cpu_ptr(hyperv_pcpu_input_arg);
+ if (!hint) {
+ local_irq_restore(flags);
+ return -ENOSPC;
+ }
+
+ hint->type = HV_EXT_MEMORY_HEAT_HINT_TYPE_COLD_DISCARD;
+ hint->reserved = 0;
+ for_each_sg(sgl, sg, nents, i) {
+ union hv_gpa_page_range *range;
+
+ range = &hint->ranges[i];
+ range->address_space = 0;
+ /* page reporting only reports 2MB pages or higher */
+ range->page.largepage = 1;
+ range->page.additional_pages =
+ (sg->length / HV_MIN_PAGE_REPORTING_LEN) - 1;
+ range->page_size = HV_GPA_PAGE_RANGE_PAGE_SIZE_2MB;
+ range->base_large_pfn =
+ page_to_hvpfn(sg_page(sg)) >> HV_MIN_PAGE_REPORTING_ORDER;
+ }
+
+ status = hv_do_rep_hypercall(HV_EXT_CALL_MEMORY_HEAT_HINT, nents, 0,
+ hint, NULL);
+ local_irq_restore(flags);
+ if ((status & HV_HYPERCALL_RESULT_MASK) != HV_STATUS_SUCCESS) {
+ pr_err("Cold memory discard hypercall failed with status %llx\n",
+ status);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+static void enable_page_reporting(void)
+{
+ int ret;
+
+ /* Essentially, validating 'PAGE_REPORTING_MIN_ORDER' is big enough. */
+ if (pageblock_order < HV_MIN_PAGE_REPORTING_ORDER) {
+ pr_debug("Cold memory discard is only supported on 2MB pages and above\n");
+ return;
+ }
+
+ if (!hv_query_ext_cap(HV_EXT_CAPABILITY_MEMORY_COLD_DISCARD_HINT)) {
+ pr_debug("Cold memory discard hint not supported by Hyper-V\n");
+ return;
+ }
+
+ BUILD_BUG_ON(PAGE_REPORTING_CAPACITY > HV_MEMORY_HINT_MAX_GPA_PAGE_RANGES);
+ dm_device.pr_dev_info.report = hv_free_page_report;
+ ret = page_reporting_register(&dm_device.pr_dev_info);
+ if (ret < 0) {
+ dm_device.pr_dev_info.report = NULL;
+ pr_err("Failed to enable cold memory discard: %d\n", ret);
+ } else {
+ pr_info("Cold memory discard hint enabled\n");
+ }
+}
+
+static void disable_page_reporting(void)
+{
+ if (dm_device.pr_dev_info.report) {
+ page_reporting_unregister(&dm_device.pr_dev_info);
+ dm_device.pr_dev_info.report = NULL;
+ }
+}
+
static int balloon_connect_vsp(struct hv_device *dev)
{
struct dm_version_request version_req;
@@ -1715,6 +1801,7 @@ static int balloon_probe(struct hv_device *dev,
if (ret != 0)
return ret;
+ enable_page_reporting();
dm_device.state = DM_INITIALIZED;
dm_device.thread =
@@ -1729,6 +1816,7 @@ static int balloon_probe(struct hv_device *dev,
probe_error:
dm_device.state = DM_INIT_ERROR;
dm_device.thread = NULL;
+ disable_page_reporting();
vmbus_close(dev->channel);
#ifdef CONFIG_MEMORY_HOTPLUG
unregister_memory_notifier(&hv_memory_nb);
@@ -1751,6 +1839,7 @@ static int balloon_remove(struct hv_device *dev)
cancel_work_sync(&dm->ha_wrk.wrk);
kthread_stop(dm->thread);
+ disable_page_reporting();
vmbus_close(dev->channel);
#ifdef CONFIG_MEMORY_HOTPLUG
unregister_memory_notifier(&hv_memory_nb);
@@ -89,7 +89,7 @@
#define HV_ACCESS_STATS BIT(8)
#define HV_DEBUGGING BIT(11)
#define HV_CPU_MANAGEMENT BIT(12)
-
+#define HV_ENABLE_EXTENDED_HYPERCALLS BIT(20)
/*
* TSC page layout.
@@ -152,11 +152,18 @@ struct ms_hyperv_tsc_page {
#define HVCALL_FLUSH_GUEST_PHYSICAL_ADDRESS_SPACE 0x00af
#define HVCALL_FLUSH_GUEST_PHYSICAL_ADDRESS_LIST 0x00b0
+/* Extended hypercalls */
+#define HV_EXT_CALL_QUERY_CAPABILITIES 0x8001
+#define HV_EXT_CALL_MEMORY_HEAT_HINT 0x8003
+
#define HV_FLUSH_ALL_PROCESSORS BIT(0)
#define HV_FLUSH_ALL_VIRTUAL_ADDRESS_SPACES BIT(1)
#define HV_FLUSH_NON_GLOBAL_MAPPINGS_ONLY BIT(2)
#define HV_FLUSH_USE_EXTENDED_RANGE_FORMAT BIT(3)
+/* Extended capability bits */
+#define HV_EXT_CAPABILITY_MEMORY_COLD_DISCARD_HINT BIT(8)
+
enum HV_GENERIC_SET_FORMAT {
HV_GENERIC_SET_SPARSE_4K,
HV_GENERIC_SET_ALL,
@@ -401,8 +408,10 @@ struct hv_guest_mapping_flush {
* by the bitwidth of "additional_pages" in union hv_gpa_page_range.
*/
#define HV_MAX_FLUSH_PAGES (2048)
+#define HV_GPA_PAGE_RANGE_PAGE_SIZE_2MB 0
+#define HV_GPA_PAGE_RANGE_PAGE_SIZE_1GB 1
-/* HvFlushGuestPhysicalAddressList hypercall */
+/* HvFlushGuestPhysicalAddressList, HvExtCallMemoryHeatHint hypercall */
union hv_gpa_page_range {
u64 address_space;
struct {
@@ -410,6 +419,12 @@ union hv_gpa_page_range {
u64 largepage:1;
u64 basepfn:52;
} page;
+ struct {
+ u64 reserved:12;
+ u64 page_size:1;
+ u64 reserved1:8;
+ u64 base_large_pfn:43;
+ };
};
/*
@@ -529,4 +544,20 @@ struct hv_set_vp_registers_input {
} element[];
} __packed;
+/*
+ * The whole argument should fit in a page to be able to pass to the hypervisor
+ * in one hypercall.
+ */
+#define HV_MEMORY_HINT_MAX_GPA_PAGE_RANGES \
+ ((HV_HYP_PAGE_SIZE - sizeof(struct hv_memory_hint)) / \
+ sizeof(union hv_gpa_page_range))
+
+/* HvExtCallMemoryHeatHint hypercall */
+#define HV_EXT_MEMORY_HEAT_HINT_TYPE_COLD_DISCARD 2
+struct hv_memory_hint {
+ u64 type:2;
+ u64 reserved:62;
+ union hv_gpa_page_range ranges[];
+} __packed;
+
#endif
@@ -27,6 +27,7 @@
struct ms_hyperv_info {
u32 features;
+ u32 priv_high;
u32 misc_features;
u32 hints;
u32 nested_features;
@@ -172,6 +173,7 @@ void hyperv_report_panic(struct pt_regs *regs, long err, bool in_die);
bool hv_is_hyperv_initialized(void);
bool hv_is_hibernation_supported(void);
void hyperv_cleanup(void);
+bool hv_query_ext_cap(u64 cap_query);
#else /* CONFIG_HYPERV */
static inline bool hv_is_hyperv_initialized(void) { return false; }
static inline bool hv_is_hibernation_supported(void) { return false; }